Optical tissue identification: weefseldifferentiatie met behulp van licht

Hele tekst

(1)

(2) ISBN: Copyright:. © 2018 G.C. Langhout, all right reserved.. De reproductie van dit proefschrift werd financieel ondersteund door: Universiteit Twente - faculteit TNW en Philips Research..

(3) OPTICAL TISSUE IDENTIFICATION WEEFSELDIFFERENTIATIE MET BEHULP VAN LICHT. PROEFSCHRIFT ter verkrijging van de graad van doctor aan de Universiteit Twente, op gezag van de rector magnificus, prof. dr. T.T.M. Palstra, volgens besluit van het College voor Promoties in het openbaar te verdedigen op vrijdag 19 oktober 2018 om 12.45 uur door Gerrit Cornelis Langhout geboren op 28 maart 1986 te Bergambacht.

(4) Dit proefschrift is goedgekeurd door de promotores; Prof. dr. T.J.M. Ruers en prof. dr. H.J.C.M Sterenborg.

(5)

(6) De promotiecommissie Voorzitter en secretaris: Prof.dr.ir. J.W.M. Hilgenkamp. Universiteit Twente, Enschede. Promotores Prof. dr. T.J.M. Ruers. Universiteit Twente, Enschede. Prof. dr. H.J.C.M. Sterenborg. Antoni van Leeuwenhoek en Amsterdam Universitair Medisch Centrum /AMC. Co-promotor: Dr. K.F.D. Kuhlmann. Antoni van Leeuwenhoek, Amsterdam. Referent: Prof. dr. A.J.M. Balm . Antoni van Leeuwenhoek, Amsterdam. Leden: Prof. dr. M.M.A.E. Claessens. Universiteit Twente, Enschede. Prof. dr. S. Manohar. Universiteit Twente, Enschede. Prof. dr. J.A. van der Hage. Leids Universitair Medisch Centrum, Leiden. Prof. dr. B.H.W. Hendriks. Philips Research, Eindhoven en Technische Universiteit Delft.

(7) Contents Summary. 10. Samenvatting Chapter 1. General introduction. 13. Chapter 2. Near-infrared fluorescence (NIRF) imaging. 25. in breast-conserving surgery: Assessing intraoperative. techniques. in. tissue-. simulating breast phantoms Chapter 3. Detection in. of. resected. melanoma human. lymph. metastases nodes. 41. by. noninvasive multispectral photoacoustic imaging Chapter 4. Differentiation of healthy and malignant. 53. tissue in colon cancer patients using optical spectroscopy: a tool for image guided surgery Chapter 5. Nerve detection during surgery: Optical spectroscopy. for. peripheral. 69. nerve. localization Chapter 6. In vivo nerve identification in head and. 81. neck surgery using diffuse reflectance spectroscopy Chapter 7. Nerve detection using optical spectroscopy,. 95. an evaluation in four different settings: in human and swine, in-vivo and post mortem Chapter 8. Optimal targeting. endobronchial lung. lesions. tool. sizes. based. on. for. 111. 3D. modeling Chapter 9. General discussion. 125. Curriculum Vitae. 137. Dankwoord. 143.

(8) Kleur is de verwondering van het licht Marc van Halsendaele.

(9) Summary.

(10) Summary This thesis describes the evaluation of optical tissue identification for surgical applications. Three promising techniques are examined. After exploration of near infra-red fluorescence imaging and photoacoustic imaging, the thesis will focus on diffuse reflectance spectroscopy (DRS) for surgical tissue identification. Chapter two describes the use of an intraoperative fluorescence camera in breast cancer phantoms. The technique provides the surgeon with an image overlay of the fluorescent contrast agent in phantoms mimicking absorption and scattering properties of human breast tissue. Important benefits and drawbacks are described. Chapter three covers the application of photoacoustic imaging in surgery. The research is focused on lymph nodes in melanoma patients. Melanin, a strong optical absorber is imaged photoacoustically. The spectral identification of both tumor and blood vessels is demonstrated in phantoms and human lymph nodes ex vivo. The photoacoustic imager used is a reflective type, with the light source incorporated in the ultrasound detector. Diffuse reflection spectroscopy is presented as an optical technique able to identify both tumor and vital surrounding structures like blood vessels and nerves. Colorectal tumors are frequently in close relation with vital structures. In oncologic rectum surgery, bladderand sexual dysfunctions are both feared and high in incidence. Chapter four describes the identification of colorectal tumor using DRS. Chapter five and six describe the identification of peripheral nerves in human during surgery. Peripheral nerves are often part of the vital structures surrounding a tumor. Ideally, image guided surgery depicts both tumor and vital surrounding tissue. Chapter five describes the identification of larger nerves as proof of principle. Chapter six is committed to the detection of smaller peripheral nerves. Optimization and validation is not necessarily executed in humans in vivo. Logistically, and patient friendly, more suited for extensive measurements are a post mortem- or animal studies. However, DRS is subject to the morphological composition and biochemical make-up of the tissue, and both will change post mortem and may differ between human and animal. Chapter seven describes the optical similarities and differences between in vivo versus post mortem and human versus swine, focused on nerve identification. In chapter eight we explore to possibilities to incorporate the DRS technique into a clinical device. We choose a bronchoscopic tool to fully utilize the small size and flexibility of the DRS optical fibers. This thesis concludes with a general discussion and outlook on a use of optical tissue identification.. 10.

(11) Samenvatting Dit proefschrift beschrijft een onderzoek naar weefselherkenning met behulp van licht voor chirurgische toepassing. Drie veelbelovende technieken worden beschreven. Na verkenning van fluorescentie beeldvorming en fotoakoestiek wordt dieper ingegaan op diffuse reflectie spectroscopie (DRS) in de operatiekamer. Hoofdstuk twee beschrijft de intra-operatieve toepassing van een camera gevoelig voor fluorescentie in het nabij infrarood, in borstkanker fantomen. Het fluorescente signaal van het contrastmiddel in fantomen met de optische eigenschappen van borstweefsel wordt als overlay gepresenteerd. Belangrijke voor- en nadelen worden besproken. Hoofdstuk drie beslaat de chirurgische toepassing van fotoakoestiek. De focus ligt op lymfklieren van melanoompatienten. Melanine, dat licht krachtig absorbeert, wordt afgebeeld. Spectrale selectie stelt ons in staat het signaal van melanine en bloed te onderscheiden in fantomen en menselijke lymfklieren buiten het lichaam. We gebruikten een hand-held apparaat met de lichtbron verwerkt in de echokop. Diffuse reflectie spectroscopie is een optische techniek welke is staat is om zowel tumor als belangrijke gezonde structuren, zoals zenuwen en bloedvaten te herkennen. Darmtumoren zijn vaak nauw verbonden met dergelijke structuren waardoor mictie- en seksuele problemen frequent voorkomen na een oncologische darmoperatie. In hoofdstuk vier wordt DRS gebruikt om tumoren van het rectum en dikke darm te identificeren. Hoofdstuk vijf en zes beschrijven het intra-operatief identificeren van perifere zenuwen in patiënten. Zenuwen zijn vaak onderdeel van de vitale structuren met een nauwe relatie tot de tumor. Intra-operatieve beeldvorming zou bij voorkeur zowel de tumor als belangrijke omliggende structuren kunnen herkennen. Hoofdstuk vijf beschrijft het herkennen van grote zenuwen als ‘proof of principle’, het volgende hoofdstuk richt zich op het herkennen van kleinere zenuwtakjes. Het optimaliseren en valideren van optische technieken hoeft niet uitsluitend in patiënten uitgevoerd te worden. Metingen post mortem of in varkens zijn met het oog op patiëntveiligheid en logistiek een alternatief. Echter, DRS is gebaseerd op verschillen in weefselsamenstelling, welke mogelijk veranderen na de dood en wellicht verschillen tussen mens en varken. Hoofdstuk zeven beschrijft optische verschillen en overeenkomsten tussen het weefsel van mens en varken, in vivo en post mortem, in het kader van zenuwherkenning. In hoofdstuk acht verkennen we de mogelijkheden om de DRS techniek te verwerken in een klinische tool. We kiezen voor toepassing binnen de bronchoscopie om de minimale afmetingen en de flexibiliteit van de DRS vezels volledig te benutten. Dit proefschrift wordt afgesloten met een discussie en vooruitzichten over het gebruik van weefselherkenning met behulp van licht.. 11.

(12)

(13) CHAPTER 1. General introduction.

(14) Introduction The general concept in surgical oncology has always been resection of the tumor while conserving healthy tissue. Surgeons focus on optimizing the balance between wide (safe) resection and maximum conservation of tissue and function. Tumor involved margins are associated with an increased risk of local recurrence for several malignancies1-5. Improving functional outcome while maintaining radical resections demands precise information on shape and position of the tumor and the relation to important surrounding structures including nerves and vessels. For this information the surgeon relies on pre-operative imaging and his own real-time observations (visual and tactile) during the procedure. Pre-operative imaging, such as computed tomography (CT) and magnetic resonance imaging (MRI) have made impressive progress over the past decades in terms of availability and image quality 6,7. For the vast majority of cases, pre-operative imaging is able to visualize both tumor and vital surrounding structures as well as their anatomical relation. On the other hand, the rise of laparoscopic techniques in the same period led to loss of tactile feedback and shifted visible feedback from direct vision to indirect camera images. Despite the progress in imaging, everyday surgery heavily relies on the visual and tactile ability of the surgeon. This is because during surgery, the changes in tissue shape compromise the relation between the preoperative images and the actual situation in the surgical field. These shape changes are caused by the tissue manipulation during surgery, requiring either continuous intraoperative imaging or adaptation of the pre-operative image to keep up with the evolving surgical procedure. Up to now, only ultrasound (US) imaging fulfills these criteria. The advantage of intra-operative ultrasound in breast cancer surgery has recently been emphasized 8 , however US focuses mainly on the tumor itself rather than the relation of the tumor with important surrounding structures7,9,10. To maximize functional outcome while maintaining radical excision, accurate identification of both tumor and vital surrounding structures is crucial. When using technical assistance. Figure 1.01. Light is electromagnetic radiation within a specific range of wavelengths. The human eye is sensitive to light with a wavelength of ~390 to ~740 nm. Presently, optical imaging is feasible using a spectral range from 300 nm to 2000 nm. Electromagnetic radiation with a wavelength shorter than visible light includes Röntgen and nuclear radiation. Microwaves and radio waves are electromagnetic radiation with a longer wavelength than visible light.. 14.

(15) to facilitate identification, the information must be up to date or even real time to cope with the continuous tissue manipulation during surgery. Image guidance for surgical application should therefore effectuate a continuous insight/overview of the position of the tumor and its relation with surrounding healthy structures. To enhance acceptance, such a technique should have minimal influence on the existing workflow. Optical imaging is considered safe, fast, inexpensive, makes use of non-ionizing radiation and enables real-time anatomical and functional imaging11. It has the potential to supply the surgeon with information about both the tumor and the surrounding structures. Therefore it promotes complete resection of otherwise challenging tumors while conserving the maximum function by meticulously avoiding damage to vital surrounding structures. The use of optical tissue identification in medical practice could include intraoperative guidance, image-guided biopsy and real time feedback in other minimal invasive procedures.. Basic principles of optical imaging Registration of photons (light) is the essence of optical imaging. As the photons travel through the tissue, the light is influenced by absorption and scattering. Where absorption attenuates the light and scattering redirects it. Both absorption and scattering differ per tissue type and depend on the color (wavelength) of the light. The wavelengths (spectral range) used for optical imaging range from ultraviolet at ~ 300nm toward infrared at ~2 µm. Optical imaging outperforms the human eye by sensitivity, broader spectral range and higher spectral resolution (i.e. the ability to differentiate between light of different wavelengths). As a reference: in contrast to the wide spectral range of optical imaging, the human eye contains only 4 different photosensitive cells of which 3 are color specific, and is sensitive to light with a wavelength of 390 to 700 nm (Fig. 1.01). Intra-operative optical techniques can be classified based on their spatial discrimination. Point measurements contain no spatial information and therefore supplies integrated values for the measured volume. Diffuse reflectance spectroscopy, as example of a point measurement, typically describes the measured volume as a whole, usually with detailed spectral information. Imaging techniques that produce an image contain spatial resolved information and therefore supply specific information for segments of the measured volume or area. A typical example for imaging is a camera, containing the relation in space to other pixels as well as (limited) information per pixel (grayscale intensity or red-green-blue values). A recent development is the hyperspectral camera that can acquire a full spectrum for each pixel. This recent technology was not introduced in the clinic at the time the research for this thesis was executed it will not be addressed here. In this thesis, three optical imaging modalities will be discussed: near infra red fluorescence imaging (NIRF), photoacoustic imaging (PAI) and diffuse reflectance spectroscopy (DRS). These techniques were selected based on the stage of research: all three have been. 15. 1.

(16) used in surgery in a research setting but did not reach broad clinical acceptance, yet. Other imaging techniques were left out the scope of this thesis because either the imaged area is extremely small compared to the surgical field or the imaging takes more time than tolerable in surgical procedures. Examples are confocal microscopy12 , two-photon microscopy12 and Optical Coherence Tomography (OCT)13. Raman spectroscopy14, a technique based on the detection of an energy shift in photons due to inelastic scattering, is also left out of scope because at the start of this work it was not available as a modality for clinical research.. Near-infra red fluorescence optical imaging An intra-operative optical imaging system produces an image of the surgical field and may add information by enhancing contrast or by visualizing characteristics invisible to the human eye. Cameras can be used in open surgery or during laparoscopy, but, in contrast to X-ray or MRI, require a clear view to the surface investigated. To enhance the optical contrast, optical contrast agents may be used, either conventional or targeted. Targeted contrast agents are engineered by molecular techniques to accumulate on a specific (target) tissue type. Other non-targeted contrast agents can visualize metabolism, like Fludeoxyglucose (FDG) in positron emission tomography (PET) imaging or anatomy, for example vascular contrast agents. In fluorescence imaging, the information is typically presented as an overlay image, ideally in real-time. In near-infrared fluorescence imaging (NIRF) the tissue is illuminated with (laser) light to excite a targeted fluorescent molecule (figure 1.02, adopted from chapter 2). After excitation, this molecule emits a photon with a longer wavelength; a different color. A highly sensitive optical system records only the fluorescent signal. Due to the relatively low absorption and scattering in the 650-900 nm ‘diagnostic window’, the system can (in optimal cases) identify fluorescence up to a centimeter deep. Auto-fluorescence (fluorescence by body molecules) in the near-infrared region is rare, enabling high sensitivity for NIRF imaging.. Figure 1.02. Overlay image from a breast phantom with a fluorescent inclusion, acquired with a multispectral camera (adopted from chapter 2).. 16.

(17) Two mayor disadvantages accompany intra-operative fluorescence imaging. First, the technique is intended to visualize a specific fluorophore, all other structures, including vital ones, are regarded as background. Although simultaneous use of a combination of differently targeted probes is suggested15 and has even performed preclinically16 , the number of distinct fluorescent probes is physically limited. This limitation is based on the spectral difference that two probes must exhibit to differ spectroscopically. Furthermore, the need to discriminate between fluorescent agents raises the technical requirements on the camera system. A second intrinsic drawback is lack of information about the depth of the origin of the signal in the tissue. In small animals this can be overcome, in some extent, by optical tomography. However, the glow effect caused by scattering of light remains a problem. This glow results in a similar image for either a weak but larger diffuse source at the surface or a strong focal source at depth (figure 1.03, adapted from chapter 2). The difference, however, is crucial in oncologic surgery as a larger superficial source requires a totally different approach than a focal source at depth. In addition to scattering, optical absorption limits the penetration depth. In fig 1.03, the hot spot of the inclusion not only blurs out, but weakens with increasing depth, due to scattering as well as absorption.. 4 mm. 7 mm. 11 mm. 15 mm. 18 mm. 21 mm. 1 cm. 1 cm Figure 1.03. Phantoms with scattering and absorption coefficients matched to breast tissue, with a 5 mm cubic fluorescent inclusion at different depths. The phantom is cast in a round glass beaker, the edge of the phantom and surrounding is cropped (set to black). The image shows that most of the spatial information is lost when the fluorescent inclusion is positioned at some depth. As a consequence, scattering hampers the discrimination between a weak superficial source and a strong focal source at depth.. Sensing optical absorption in 3D With photoacoustic imaging (PAI), two disadvantages of NIRF (no identification of the vital surrounding structures and the lack of information about the depth of the signal) can be overcome. PAI images optical absorption in 3 dimensions and can therefore discriminate 17. 1.

(18) between superficial and deeper lying sources17. The technique uses pulsed light and is based on tissue specific absorption. The absorption of pulsed light results in a pulsed thermal expansion. In case of a very short pulse (~1 nsec), the rapid expansion causes a sound wave that can be detected using an ultrasound detector (Figure 1.04). The location of the absorbing structure can be determined analog to conventional sonography. As in fluorescent imaging, the result is an overview image with the contrast highlighted. However, there are two important differences. First, instead of a fluorescent emission, the tissue specific chromophores emit ultrasound waves. Sound waves are not influenced by optical scattering and can be traced to the source in 3D. As ultrasound waves are less affected by scattering compared to light, PAI localizes deeper lying sources more precise compared to NIRF imaging. Second, PAI can discriminate between different optical absorbers when sequentially illuminating the tissue with a variety of wavelengths (multispectral PAI). Doing so, simultaneous visualization of multiple body chromophores as well as optical contrast agents is possible. Multispectral PAI should be able to visualize and recognize both tumor and vital structures like blood vessels in 3D images, with minimal distortion due to optical scattering. Nevertheless, photoacoustic imaging is a complex technique investigated by few research groups. PAI is time-consuming especially in multispectral mode and requires acoustic tissue contact during imaging.. illumination with pulsed laser light. transducer. absorption causes termal expansion. transducer. the expansion creates an ultrasound wave. transducer. the wave is detected using a transducer. transducer. Figure 1.04. Schematic representation of photoacoustic imaging; an ultrasound wave is elicited by thermal expansion after absorption of pulsed laser light. In the visualized setup, the laser light source is separated from the ultrasound array.. Optical spectroscopy In optical spectroscopy, tissue discrimination is usually based on intrinsic spectral differences, without the use of contrast agents. In general, a continuous spectrum is recorded from which (multiple) discriminating parameters are derived. Within optical spectroscopy, several different techniques can be distinguished such as Diffuse Reflectance Spectroscopy (DRS) and Fluorescence Spectroscopy (FS). With DRS, tissue is illuminated with white light and the reflectance is collected, often using a fiber optic cable containing fibers of 50-200 µm in diameter. When the light travels from the emission fiber to the collection fiber, the photons interact with the tissue. The spectrum recorded from the collection fiber differs from the spectrum of the light source as the light is scattered and partially absorbed by the tissue. Both scattering and absorption are tissue specific and depend on the wavelength of the 18.

(19) incident light18 . In this way, different chromophores in the tissue such as hemoglobin (oxygenated and deoxygenated), b-carotene, water, lipids, and collagen can be recognized and quantified based on a characteristic spectrum. Fluorescence spectroscopy (FS) is the recording of the fluorescence signal after (laser) excitation. FS adds the possibility to determine intrinsic fluorophores in the measured tissue, such as collagen, elastin, FAD, NADH, and porphyrins18 . Collagen and elastin are structural proteins and are associated with tissue structure, whereas NADH and FAD levels are indicative for cellular energy metabolism. DRS-FS has been shown to identify tumors of various origins19-22 . DRS is a point measurement technique and therefore does not provide an (overview) image. Furthermore, it requires contact to the tissue of interest, which in some situations can be a limitation.. Surgical application The optical techniques described have different technical requirements when applied in the clinic. NIRF imaging needs a clear, non scattering medium like air or water between the tissue and the camera whereas PAI and DRS require tissue contact. Also, the space required for the setup differs. PAI needs the most space with a laser source and ultrasound array. In theory the technology could be scaled down to integration into laparoscopic instruments, but a laparoscopic instrument capable of photoacoustic imaging has not been presented so far. Laparoscopic NIRF cameras are commercially available 23. NIRF during surgery can therefore be done either in open surgery or during laparoscopic procedures. For application of DRS during surgery the small diameter of the optical fibers in DRS make incorporation into all kind of devices possible. Spliethoff et al. described a biopsy tool with optical fibers for DRS fully incorporated 24. Rathmell et al. presented a needle for epidural anesthesia equipped with optical fibers 25. Eventually, DRS could be used during the surgical procedure by incorporating optical fibers into a smart surgical instrument.. Clinical device Intra-operative cameras in clinical research are often already used as intended for clinical application: Placed in the perspective of the surgeon and delivering an overview image. For photoacoustic imaging, the first clinical application will probably be integrated into a hand held ultrasound probe, as used in chapter 3. However, for DRS the clinical application is less clear. Although the technique is often presented as a needle or stylus, smart optical devices for specific clinical applications were presented. Spliethoff et al. used smart optical biopsy tool for transthoracal lung biopsies, with optical fibers fully incorporated into a functional biopsy tool 24. Rathmell et al. presented a functional needle for epidural anesthesia equipped with optical fibers 25. The small dimensions of the optical fibers in DRS make incorporation into existing clinical devices possible; as demonstrated by the two examples above. However, clinical. 19. 1.

(20) application of optical spectroscopy should not be limited to spicing up existing devices. In chapter eight, we explore the dimensional requirements of a bronchoscopic biopsy tool.. Goal of this thesis For optimal results in surgical treatment, accurate localization and orientation of both tumor and vital surrounding tissue is crucial to achieve radical resection while conserving maximum function and physical appearance. The goal of this thesis is to evaluate three different techniques for intraoperative imaging on their ability to identify both tumor and surrounding vital structures, the dependency on other systems for overview and orientation, and the constraints for integration of these techniques in the clinical work-flow.. Outline of this thesis Chapter two describes the use of an intraoperative fluorescence camera in breast cancer phantoms. The technique provides the surgeon with an image overlay of the fluorescent contrast agent in phantoms mimicking absorption and scattering properties of human breast tissue. In this chapter the penetration depth or depth of view of NIRF imaging is assessed by imaging a contrast rich inclusion in a tissue mimicking phantom at multiple depths. The relation between the depth of the signal and the recorded pattern at the surface is discussed. The integration into surgical practice is evaluated by performing lumpectomy in breast cancer phantoms containing fluorescence inclusions. In chapter three the possibilities for photoacoustic imaging in surgery are investigated. In PAI, ultrasound waves are detected that were caused by optical absorption in the tissue. Therefore PAI does not suffer from optical scattering as in NIRF imaging. Furthermore, it provides 3D imaging and thus also providing information about the depth of the tissue of interest. The experimental work is focused on diagnosing positive lymph nodes in melanoma patients. Tumor positive lymph nodes contain both tumor deposits as well as blood vessel. This makes it a suitable test case to detect both tumor and vital surrounding structures. The spectral identification of both tumor and blood vessels is demonstrated in phantoms and human lymph nodes ex vivo. Furthermore, the penetration depth of the technique is assessed. In oncologic colorectal surgery, tumor is often in close relation to vital structures. The close relation of peripheral nerves with the surgical field may result in nerve damage, leading to bladder- and sexual dysfunction. Image guidance in colorectal surgery should preferably identify malignancy and healthy surrounding tissue including peripheral nerves. Chapter four describes the identification of colorectal tumors using diffuse reflection spectroscopy, as optical technique able to identify both tumor and vital surrounding structures. In an ex vivo model (resection specimen), tissue from colorectal cancer is distinguished from multiple healthy surrounding tissues. Chapter five and six describe the use of DRS for the identification of peripheral nerves 20.

(21) in human subjects during surgery. Peripheral nerves are often part of the vital structures surrounding a tumor. Chapter five describes the identification of larger nerves (femoral or sciatic nerve) as a proof of principle, chapter six is committed to the detection of smaller peripheral nerves (the greater auricular nerve, the accessory nerve or the facial nerve) by DRS. In chapter seven, optical differences between human in vivo and alternative models for optimization and validation of optical techniques are evaluated. To move the technique from a strictly controlled experimental setting into a mature clinically applied device is a long process that will require multiple optimization and validation cycles. Not all these cycles need necessarily to be executed in humans in vivo. For example, for extensive measurements either an ex vivo post mortem setup or an animal model may be more practical. However, DRS is subject to the morphological composition and biochemical make-up of tissue. Both these properties will change post mortem and may differ between human and animal. In chapter seven we compared macroscopy, histology and DRS spectra between human in vivo and alternative models for nerve and surrounding tissue. The results contribute to decision making in model selection for future optimization and validation cycles. In chapter eight, a possible application of optical tissue identification is further elaborated. The small size of the optical fibers in DRS supports the development of innovative tools. The bronchial tree is an anatomical location where a smaller diagnostic device, such as a bronchoscopic biopsy tool, could reach more (peripheral) lesions. By the exchange of the conventional visual feedback in bronchoscopy for optical tissue identification, the size of intra-bronchiolar tools could be drastically decreased. Based on simulations of the bronchial three and analysis of patient CT scans, the relation between the size of a bronchoscopic tool and the percentage of reachable lesions is defined and we estimate the optimal size of bronchoscopic biopsy tools to reach also peripheral lung lesion for biopsy. This thesis is concluded with a general discussion in chapter nine featuring an outlook on the use of optical tissue identification for surgical application.. 21. 1.

(22) References 1.. Quirke P, Dixon M, Durdey P, Williams N. Local recurrence of rectal adenocarcinoma due to inadequate surgical resection: histopathological study of lateral tumor spread and surgical excision. The Lancet 1986; 328(8514):996-999.. 2.. Connolly JA, Shinohara K, Presti JC, Carroll PR. Local recurrence after radical prostatectomy: characteristics in size, location, and relationship to prostate-specific antigen and surgical margins. Urology 1996; 47(2):225-231.. 3.. Schnitt SJ, Abner A, Gelman R, Connolly JL, Recht A, Duda RB, Eberlein TJ, Mayzel K, Silver B, Harris JR. The relationship between microscopic margins of resection and the risk of local recurrence in patients with breast cancer treated with breastconserving surgery and radiation therapy. Cancer 1994; 74(6):1746-1751.. 4.. Thomas JM, Newton-Bishop J, A’hern R, Coombes G, Timmons M, Evans J, Cook M, Theaker J, Fallowfield M, O’neill T. Excision margins in high-risk malignant melanoma. New England Journal of Medicine 2004; 350(8):757-766.. 5.. Gerrand C, Wunder J, Kandel R, O’Sullivan B, Catton C, Bell R, Griffin A, Davis A. Classification of positive margins after resection of soft-tissue sarcoma of the limb predicts the risk of local recurrence. Bone & Joint Journal 2001; 83(8):1149-1155.. 6.. Rutt BK, Lee DH. The impact of field strength on image quality in MRI. Journal of Magnetic Resonance Imaging 1996; 6(1):57-62.. 7.. Becker CR, Ohnesorge BM, Schoepf UJ, Reiser MF. Current development of cardiac imaging with multidetector-row CT. European journal of radiology 2000; 36(2):97-103.. 8.. Haloua MH, Volders JH, Krekel NM, Cardozo AML, de Roos WK, de Widt-Levert LM, van der Veen H, Rijna H, Bergers E, Jóźwiak K. Intraoperative ultrasound guidance in breast-conserving surgery improves cosmetic outcomes and patient satisfaction: results of a multicenter randomized controlled trial (COBALT). Annals of surgical oncology 2016; 23(1):30-37.. 9.. Gray RJ, Pockaj BA, Garvey E, Blair S. Intraoperative Margin Management in Breast-Conserving Surgery: A Systematic Review of the Literature. Annals of surgical oncology 2017:1-10.. 10. Hiramoto JS, Feldstein VA, LaBerge JM, Norton JA. Intraoperative ultrasound and preoperative localization detects all occult insulinomas. Archives of surgery 2001; 136(9):1020-1026. 11. Keereweer S, Kerrebijn JD, Van Driel PB, Xie B, Kaijzel EL, Snoeks TJ, Que I, Hutteman M, Van der Vorst JR, Mieog JSD. Optical image-guided surgery—where do we stand? Molecular Imaging and Biology 2011; 13(2):199-207. 12. Diaspro A. Confocal and two-photon microscopy: foundations, applications and advances. Confocal and Two-Photon Microscopy: Foundations, Applications and Advances, by Alberto Diaspro (Editor), pp 576 ISBN 0-471-40920-0 Wiley-VCH, November 2001 2001:576. 13. Huang D, Swanson EA, Lin CP, Schuman JS, Stinson WG, Chang W, Hee MR, Flotte T, Gregory K, Puliafito CA. Optical coherence tomography. Science (New York, NY) 1991; 254(5035):1178. 14. Colthup N. Introduction to infrared and Raman spectroscopy: Elsevier. 2012. 15. Nguyen QT, Tsien RY. Fluorescence-guided surgery with live molecular navigation [mdash] a new cutting edge. Nature reviews cancer 2013; 13(9):653-662. 16. Fernández A, Vendrell M. Smart fluorescent probes for imaging macrophage activity. Chemical Society Reviews 2016; 45(5):1182-1196. 17. Xu M, Wang LV. Photoacoustic imaging in biomedicine. Review of scientific instruments 2006; 77(4):041101. 18. Cheong W-F, Prahl SA, Welch AJ. A review of the optical properties of biological tissues. IEEE journal of quantum electronics 1990; 26(12):2166-2185. 19. Bard MP, Amelink A, Skurichina M, Noordhoek Hegt V, Duin RP, Sterenborg HJ, Hoogsteden HC, Aerts JG. Optical spectroscopy for the classification of malignant lesions of the bronchial tree. Chest 2006; 129(4):995-1001. 20. Zhou C, Choe R, Shah N, Durduran T, Yu G, Durkin A, Hsiang D, Mehta R, Butler J, Cerussi A. Diffuse optical monitoring of blood flow and oxygenation in human 22.

(23) breast cancer during early stages of neoadjuvant chemotherapy. J Biomed Opt 2007; 12(5):051903-051903-051911. 21. Wang H-W, Jiang J-K, Lin C-H, Lin J-K, Huang G-J, Yu J-S. Diffuse reflectance spectroscopy detects increased hemoglobin concentration and decreased oxygenation during colon carcinogenesis from normal to malignant tumors. Optics Express 2009; 17(4):2805-2817. 22. Brown JQ, Bydlon TM, Richards LM, Yu B, Kennedy SA, Geradts J, Wilke LG, Junker MK, Gallagher J, Barry WT. Optical assesssment of tumor resection margins in the breast. Selected Topics in Quantum Electronics, IEEE Journal of 2010; 16(3):530-544. 23. van den Bos J, Schols RM, Luyer MD, van Dam RM, Vahrmeijer AL, Meijerink WJ, Gobardhan PD, van Dam GM, Bouvy ND, Stassen LP. Near-infrared fluorescence cholangiography assisted laparoscopic cholecystectomy versus conventional laparoscopic cholecystectomy (FALCON trial): study protocol for a multicentre randomised controlled trial. BMJ open 2016; 6(8):e011668. 24. Spliethoff JW, Prevoo W, Meier MA, de Jong J, Klomp HM, Evers DJ, Sterenborg HJ, Lucassen GW, Hendriks BH, Ruers TJ. Real-time in vivo tissue characterization with diffuse reflectance spectroscopy during transthoracic lung biopsy: a clinical feasibility study. Clinical cancer research 2016; 22(2):357-365. 25. Rathmell JP, Desjardins AE, van der Voort M, Hendriks BH, Nachabe R, Roggeveen S, Babic D, Söderman M, Brynolf M, Holmström B. Identification of the Epidural Space with Optical SpectroscopyAnIn Vivo Swine Study. The Journal of the American Society of Anesthesiologists 2010; 113(6):1406-1418.. 23.

(24)

(25) CHAPTER 2. Near-infrared fluorescence (NIRF) imaging in breast-conserving surgery: Assessing intraoperative techniques in tissue-simulating breast phantoms. Breast-conserving surgery (BCS) results in tumor-positive surgical margins in up to 40% of the patients. Therefore, new imaging techniques are needed that support the surgeon with real-time feedback on tumor location and margin status. In this study, the potential of near-infrared fluorescence (NIRF) imaging in BCS for pre- and intraoperative tumor localization, margin status assessment and detection of residual disease was assessed in tissue-simulating breast phantoms. Breast-shaped phantoms were produced with optical properties that closely match those of normal breast tissue. Fluorescent tumor-like inclusions containing indocyanine green (ICG) were positioned at predefined locations in the phantoms to allow for simulation of (i) preoperative tumor localization, (ii) real-time NIRF-guided tumor resection, and (iii) intraoperative margin assessment. Optical imaging was performed using a custom-made clinical prototype NIRF intraoperative camera. Tumor-like inclusions in breast phantoms could be detected up to a depth of 21 mm using a NIRF intraoperative camera system. Real-time NIRF-guided resection of tumor-like inclusions proved feasible. Moreover, intraoperative NIRF imaging reliably detected residual disease in case of inadequate resection. We evaluated the potential of NIRF imaging applications for BCS. The clinical setting was simulated by exploiting tissue-like breast phantoms with fluorescent tumor-like agarose inclusions. From this evaluation, we conclude that intraoperative NIRF imaging is feasible and may improve BCS by providing the surgeon with imaging information on tumor location, margin status, and presence of residual disease in real-time. Clinical studies are needed to further validate these results.. European Journal of Surgical Oncology R.G. Pleijhuis, G.C. Langhout, W. Helfrich, G. Themelis, A. Sarantopoulos, L.M.A. Crane, N.J. Harlaar, J.S. de Jong, V. Ntziachristos, G.M. van Dam.

(26) Introduction Breast cancer is the most frequent malignancy in women worldwide with an estimated 1.4 million new cases in 20101. Breast-conserving therapy (BCT), consisting of breastconserving surgery (BCS) followed by radiation therapy, has become the standard treatment for T1–T2 breast tumors and is generally regarded as sufficient for this subset of patients 2 . Unfortunately, a majority of studies on the surgical margin status after BCS have shown that positive margins are detected in 20–40 % of patients, necessitating additional surgical intervention or radiotherapy 3. Two major points for improving outcome after BCS involve (i) a more reliable intraoperative tumor localization and (ii) improved real-time feedback on the presence of possible residual disease during or after excision of the tumor4. Intraoperative application of an optical imaging technique known as near-infrared fluorescence (NIRF) imaging may improve the clinical outcome of BCS 3,5. Near-infrared fluorescence imaging In recent years, significant progress has been made in the development of optical imaging systems and fluorophores for clinical applications 6,7. Several animal 5,8-10 and clinical11-15 studies have shown the potential clinical use of NIRF imaging to improve the therapeutic outcome of surgery. Compared to light in the visible spectral range (400–650 nm), application of near-infrared (NIR) light minimizes absorption by physiologically abundant molecules such as hemoglobin and lipids, which increases penetration depth16,17. Additionally, autofluorescence (the intrinsic fluorescence signal present in all living cells due to various normal metabolites and tissue constituents) is strongly reduced in the NIR spectral range. Taken together, these aspects of NIR light make it particularly suitable for use in intraoperative optical imaging applications. However, clinical application of NIRF imaging in BCS is currently limited to the non-specific intraoperative detection of the sentinel lymph node11,12,14,18-20. Tumor-targeted near-infrared fluorophores With the introduction of clinical grade tumor-targeted NIR fluorophores, NIRF imaging may be extended towards the intraoperative detection of the primary tumor 10. Several target molecules have been identified for breast cancer that may be of value for such an approach, including Her2/neu receptor 9,21,22 , vascular endothelial growth factor (VEGF) receptor 23,24, endothelial growth factor (EGF) receptor 25 and folate receptor-α 26. In tumor-targeted NIRF imaging, a tumor-targeted NIR fluorophore is administered several hours or days prior to the imaging procedure. Subsequently, an external laser is used to irradiate the breast with light in the NIR spectral range (650–900 nm)17. Upon excitation, the fluorophore will release photons of a higher NIR wavelength. Because NIR light is invisible to the naked eye, a dedicated optical imaging system is necessary to capture the NIR signal from the surgical field and digitally convert it to a visible image. Recently, we 26.

(27) and our co-workers developed a multispectral NIRF intraoperative camera system that is suitable for intraoperative use with NIR fluorophores 27. Simulation of NIRF-guided breast-conserving surgery In the current preclinical study, we evaluated intraoperative NIRF imaging applications in a simulated clinical setting as a step-up toward NIRF-guided BCS. To this end, we used tissue-simulating gelatin-based breast phantoms that mimic the optical properties of normal breast tissue 28,29. Tumor-like fluorescent inclusions of different size and shape were positioned at predefined sites in the phantoms, allowing for simulation of (i) preoperative tumor localization, (ii) real-time NIRF-guided tumor resection and (iii) intraoperative macroscopic margin assessment. The tumor-like inclusions contain the non-specific NIR fluorophore indocyanine green (ICG), to simulate for the use of tumortargeted near-infrared fluorophores in BCS. Currently, ICG (absorption and emission maximum at ~780 nm and ~820 nm, respectively) is one of the few FDA-approved NIR fluorophores available for clinical use 9. Sevick-Muraca et al. have previously shown the feasibility of NIRF imaging following microdose administration of ICG12 . Although ICG in itself is non-specific, their findings suggest that comparable microdose concentrations can be used to label cancer cells with tumor-targeted NIR fluorophores for intraoperative NIRF imaging. Importantly, new fluorophores in the NIR spectral range are currently being developed, e.g. IRDye ® 800CW, with properties more promising for intraoperative use compared to ICG 25.. Material and methods Assessment of ICG fluorescence self-quenching in agarose Because increasing concentrations of ICG may not correspond to an increased fluorescence signal due to self-quenching of ICG, different concentrations of ICG in agarose were evaluated for fluorescence activity 29,30. Briefly, an ICG stock solution was serially diluted in 10 ml sterile water (ranging from 0.5 uM to 350 uM ICG), after which 2 % agarose was added. The mixture was then heated to 70 °C and stirred until the agarose was completely dissolved. After solidification of the agarose mixture for 15 min at 4 °C, NIRF epi-illumination imaging was performed to determine maximum photon counts/ sec (settings: exposure time: 1000 ms, excitation: 780 nm, emission: 820 nm). Assessment maximal penetration depth of ICG fluorescence In order to determine the maximal penetration depth of the NIRF signal, a cubic fluorescent inclusion of 5 × 5 × 5 mm containing 14 μM ICG was positioned in phantom tissue at a depth of 30 mm. Subsequently, the surgeon excised 3–4 mm layers of phantom tissue towards 27. 2.

(28) the inclusion (remaining depths were 27, 24, 21, 18, 15, 11, 7, and 4 mm, respectively). At all depths, NIRF epi-illumination imaging was performed with the intraoperative NIRF camera system (exposure time: 3000–60000 ms, excitation 780 nm, emission 820 nm, binning: small-medium). Maximum photon counts per second exposure time were calculated as well as the full width at half maximum (FWHM) of the fluorescence signal. The FWHM is a measure for scattering and indicates the diameter of the fluorescence signal when the intensity of the signal is reduced to half the maximum. Scattering both contributes to signal loss and loss in resolution. The FWHM indicates the minimal distance between two distinct sources to be recognized as separate. Tissue-simulating breast phantoms Composition of the tissue-simulating gelatin-based breast phantoms was aimed at obtaining uniform optical properties that closely match the optical characteristics of normal breast tissue, as described in detail before 29. Additionally, the breast phantoms mimic the elastic properties of human tissue 31. Briefly, 10 % gelatin 250 (Natural Spices, Watergang, the Netherlands) was dissolved in 1 l TBS (50 mmol Tris–HCl, 150 mmol NaCl, pH 7.4). To remove molecular oxygen and prevent microbial infection, 15 mmol NaN 3 (Merck, Darmstadt, Germany) was added. The gelatin slurry was completely dissolved by heating to 50 °C and subsequently cooled down to 35 °C and maintained at this temperature. Under constant stirring, 170 μmol hemoglobin (Sigma–Aldrich, Zwijndrecht, The Netherlands) and 1 % Intralipid ® 20 % (Sigma–Aldrich) were added. Next, the gelatin mixture was poured in a custom-made pre-chilled breastshaped mold (end volume: 500 ml) to a level that corresponded to the predefined depth of the agarose inclusion. After solidification for 30 min at 4 °C, a tumor-like NIR fluorescent agarose inclusion was positioned on the surface and temporarily fixed with a small needle. Next, the remaining of the warmed gelatin mixture was added to fill up the remaining mold volume, allowing for adherence of both layers. The phantom was then stored in the dark to solidify for another 30 min at 4 °C, after which it was gently removed from its mold. In total, 4 breast phantoms were constructed with tumor-like NIR fluorescent agarose inclusions of different size and shape (Figure 2.01A) positioned at predefined depths. Imaging of the phantoms was performed directly after production of all 4 phantoms. Breast phantom #1 contained 2 similar-sized (Ø1.0 cm) sphere-shaped agarose inclusions at different depths (2.0 and 4.0 cm). Phantom #2 contained 2 sphere-shaped inclusions at the same depth (1.5 cm), differing in size (Ø0.5 cm and Ø2.0 cm). Phantom #3 contained 1 sphere-shaped (Ø1.0 cm) and 1 prolate sphere-shaped (Ø1.0 cm) agarose inclusion at the same depth (1.5 cm). Finally, phantom #4 contained 2 irregular shaped agarose inclusions of similar size (Ø1.5 cm) at different depths (1.5 and 3.0 cm).. 28.

(29) Tumor-like NIR fluorescent agarose inclusions For tumor-like NIR fluorescent agarose inclusions, 2 % agarose (Hispanagar, Burgos, Spain) was used instead of 10 % gelatin. Agarose has a higher melting point which prevents the inclusions from dissolving and leaking ICG (see below) during and after the positioning procedure in the gelatin phantom. In short, a 2 % (W/V) agarose slurry was heated to 70 °C and stirred until the agarose was completely dissolved. Subsequently, ICG (ICG-PULSION ® ; Pulsion Medical Systems, Munich, Germany) was dissolved to a final concentration of 14 μM. Finally, in order to resemble the optical appearance of the surrounding breast phantom tissue, 170 μM hemoglobin, 15 mM NaN 3 and 1 % Intralipid ® 20% were added to the tumor-like fluorescent inclusions. Tumor-like fluorescent inclusions of different size (range: 0.5 cm–2.0 cm) and shape (prolate sphere, sphere and irregular shape, Fig. 2.01A) were produced. The inclusions were integrated in the breast phantoms as indicated and chilled in the dark for 30 min at 4 °C. Imaging of each individual breast phantom was performed within 6 h. Near-infrared fluorescence imaging system A custom-made NIRF camera system was developed in collaboration with SurgOptix Inc (SurgOptix Inc, Redwood Shores, CA, USA) for real-time intraoperative imaging. The system implements a correction scheme that improves the accuracy of epi-illumination fluorescence images for light intensity variations in tissue. Implementation is based on the use of three cameras operating in parallel. The camera is mounted on a five degrees of freedom bracket. Additionally, a sixth degree (rotation) can be performed digitally. The camera allows for simultaneous acquisition of color videos and normalized fluorescence images in real-time, yielding a lateral resolution up to 66.58 μm and a variable field of view (FOV) of 13.5 W × 11 H to 115 W × 95 H (mm). A description in full detail is provided by Themelis et al 27. The invisible NIRF imaging signal was digitally converted into a pseudocolor and superposed on a color video image of the operating field, allowing for real-time, intraoperative anatomical positioning of the fluorescence signal. Simulation of intraoperative NIRF imaging Breast phantoms with tumor-like NIR fluorescent agarose inclusions were used to simulate and evaluate the potential of NIRF imaging applications in BCS (Fig. 2.01 and Fig. 2.03). In all phantoms, the location of the tumor-like fluorescent inclusions was assessed preoperatively with non-invasive NIRF imaging. In phantoms #1 and #2, the tumor-like fluorescent inclusions were subsequently excised using conventional surgical equipment, guided solely by visual inspection, tactile information, and preoperatively obtained NIRF imaging data. The surgeon was asked to indicate when he believed a complete excision of the tumor was reached. Subsequently, the NIRF camera system was applied to assess the feasibility of NIRF-guided macroscopic margin assessment of the surgical cavity and. 29. 2.

(30) A Inclusions. B NIRF Camera system. C Breast phantom. Figure 2.01. | Fluorescent tumor-like agarose inclusions differing in size and shape (AI–III) were integrated in breast-shaped phantoms (CI) prior to surgery. Preoperatively, the location of the tumor-like inclusion was assessed non-invasively using a NIRF camera system (B). Intraoperatively, the inclusion was excised under real-time NIRF guidance or guided solely by visual and tactile information (CII). At the end of the surgical procedure, the NIRF camera system was applied to inspect for residual disease and evaluate the extend of surgery (CIII).. excised tissue fragments. In case of an incomplete excision, the surgeon was asked to perform a re-excision under real-time NIRF guidance. In phantoms #3 and #4, the tumor-like fluorescent inclusions were localized and excised under real-time NIRF guidance. While approaching the tumor-like fluorescent inclusions, the surgeon was supported with both visible and audible information. In short, the detected fluorescence signal was depicted on a TFT-screen and was made quantitatively audible using a digitally generated sound-pitch. In this approach, an increase in sound-pitch represents an increase in fluorescence signal indicating the approximation of a tumor-like fluorescent inclusion.. 30.

(31) Results ICG fluorescence self-quenching in agarose To determine the optimal ICG concentration in agarose, a serial range of increasing ICG concentrations was analyzed for fluorescence characteristics. The optimal fluorescence signal was observed at a concentration of approximately 10 μM ICG (Fig. 2.04). These results are comparable to self-quenching characteristics of ICG as previously determined in gelatin 29. Maximal penetration depth of ICG fluorescence The maximal tissue penetration depth of a detectable fluorescent ICG inclusion was reached at a depth of 21 mm (Fig. 2.02C).. A Fluorescence signal intensity. C NIRF penetration depth. B Scattering. 4 mm. 7 mm. 11 mm. 15 mm. 18 mm. 21 mm. Figure 2.02. | Fluorescence signal intensity related to depth-location in tissue-like phantoms is shown for fluorescent agarose inclusions placed at varying depths in phantom tissue (A). Depth (mm) of the inclusion and maximum photon counts per second exposure time are depicted on the horizontal and vertical axis, respectively. Moreover, scattering of the fluorescence signal is shown (B), with depth (mm) and full width at half maximum (pixels) on the horizontal and vertical axis, respectively. For determination of the NIRF signal penetration depth with the NIRF intraoperative camera system (C), the surgeon repeatedly excised 3–4 mm tissue layers, working his way towards a fluorescent inclusion placed at 30 mm depth in breast phantom tissue. At 30, 27 and 24 mm inclusion depth, no NIRF signal could be detected (not shown). Images were corrected (normalized) for an exposure time of 1000 ms.time NIRF guidance or guided solely by visual and tactile information (CII). At the end of the surgical procedure, the NIRF camera system was applied to inspect for residual disease and evaluate the extend of surgery (CIII).. 31. 2.

(32) Simulation of intraoperative NIRF imaging Preoperative NIRF-guided localization of tumor-like fluorescent agarose inclusions was performed in 4 different breast phantoms. The various tumor-like inclusions positioned at a depth of ≤2.0 cm were detectable with the NIRF camera system (Fig. 2.03A-I). Tumor-like inclusions positioned at depths of 3.0 and 4.0 cm could not be detected preoperatively. Tumor-like inclusions in phantom #1 and #2 were excised using conventional techniques for tumor localization. In phantom #1, one out of two tumor-like inclusions proved to be only partially excised, as evidenced by a remnant strong fluorescence signal in the surgical cavity detected by the NIRF camera system (Fig. 2.03C-III). In phantom #2, the excision of one out of two inclusions was found to be incomplete. In case of residual fluorescence, the surgeon could detect and excise (theranostic procedure) the remnant inclusion under real-time NIRF guidance (Figure 3C-IV and supplemental video 1). In all cases, NIRF-guided re-excision resulted in a complete excision, without the need for additional excision of large breast phantom fragments (Fig. 2.03C-V).. Near-Infrared Fluorescence (NIRF) imaging applications A Preoperative tumour localization. C NIRF guided surgery. B Intraoperative tumour localization. Figure 2.03. | Overview of NIRF applications in breast-conserving surgery. In the case of relatively superficial lesions (≤2 cm), NIRF allows for preoperative localization of fluorescent tumor-like inclusions (A). Intraoperative NIRF imaging guides the surgeon towards the tumor-like agarose inclusions (B + C) and allows for intraoperative assessment of surgical margin status (CII–V) and detection of residual disease (CIII). Color bars next to the color overlay indicate a threshold at 800 counts. Pixels with values above the threshold were superposed on the color video (overlay). Exposure time was set to 150 ms for all images.. 32.

(33) In phantoms #3 and #4, the tumor-like inclusions were located (Fig. 2.03B) and excised (Fig. 2.03C) under real-time NIRF guidance. Although the inclusion at 3.0 cm depth in phantom #4 could not be detected preoperatively (Fig 2.03B-I), it was detectable using the NIRF camera system after an incision of approximately 1 cm of superficial phantom tissue (Fig. 2.03B-II). In phantom #3, no residual tumor-like inclusion material could be detected after initial NIRF-guided excision, while the excision of one out of two irregular inclusions in phantom #4 was found to be incomplete. Again, subsequent NIRF-guided re-excision resulted in a complete excision. During the surgical simulation procedure, the approximation of the surgeon towards tumor-like fluorescent agarose inclusions was guided by both visual information on a TFT-screen and audible sound-pitched information. The approach resulted in a clear change in the signal strength of the fluorescence image that was accompanied with an increase of the sound pitch at ~15 mm prior to excision of the tumor-like agarose inclusion. These signals assisted the surgeon in carefully advancing the margins.. Photon counts (x1000). 14. ICG signal optimum. 12 10 8 6 4 2 0 −1 10. 0. 1. 2. 10 10 10 Concentration ICG (μM). 10. 3. Figure 2.04. | Optimal ICG concentration in 2% agarose was determined in breast phantom tissue. Seventeen different concentrations (each 10 ml, ranging from 0.5 μM to 350 μM) were imaged at once with the intraoperative NIRF camera system. Working distance 45 cm, field of view 160 W × 130 H (mm), exposure time: 1000 ms.. 33. 2.

(34) Discussion Tissue-like phantoms and tumor-like inclusions The composition of the breast phantoms was based on data published by De Grand et al., who developed and validated phantoms to mimic the basic optical characteristics (absorption and scattering coefficients) of breast tissue 29. The absorption of photons by both cellular organelles and blood was simulated by hemoglobin, which gives the phantoms a deep red color 32,33. Additionally, Intralipid ® was added to mimic scattering properties of breast tissue 28 . In order to resemble the clinical situation as close as possible, tumor-like fluorescent agarose inclusions were incorporated in the breast phantoms. The agarose-based inclusions simulate the firm-elastic consistency of tumor tissue and allow for surgical margin status assessment, both intraoperatively (NIRF-guided surgery) and ex vivo (NIRF-guided macroscopic margin assessment). The relatively low concentration of ICG used in this study resembles the potential application of microdose tumor-targeted fluorophores (ranging from 1 μM to 100 μM) in BCS. Although the phantoms used in this study are homogeneous, and therefore do not possess the complex structures that characterize mammary tissue, they do provide a tool for assessing the value of theoretical assumptions and indicate generally important features of future clinical NIRF imaging applications in BCS. Near-infrared fluorescence imaging: strengths and drawbacks NIRF imaging offers a promising technique for real-time NIRF-guided surgery in BCS with little interference in the standard surgical procedure or changes in the design of the operating theater. The technology is considered safe, fast, makes use of non-ionizing radiation, and has a high resolution 3,10. However, NIRF imaging does have limitations originating from the intrinsic characteristics of light propagation through tissue, including scattering and absorbance17. Additionally, due to limited depth resolution and a non-linear dependence of the signal detected and the depth of the fluorescence activity, NIRF imaging by epi-illumination with our current camera system seems of limited value for preoperative localization of tumors. This applies in particular to situations where the tumor is located relatively deep (>2 cm) in fat and glandular tissue of the breast. However, since the surgeon, by definition, will bring the area of interest closer to the surface during surgery, our multispectral NIRF camera system is well-suited for intraoperative imaging applications. NIRF imaging instruments designed for preoperative imaging, e.g. the SoftScan (ART, Advanced Research Technologies, Saint-Laurent (Quebec), Canada), show penetration depths far superior (up to 15 cm) to intraoperative imaging systems 34. This difference in. 34.

(35) penetration depth is due to the application of different imaging strategies which are largely incompatible with surgery, e.g. trans-illumination and the need for light-conducting liquid media. Non-specific versus tumor-targeted fluorophores Several possibilities exist for delivering fluorophores to the tumor. One possibility would be to inject a non-specific fluorophore (e.g. ICG) into the tumor under stereotactic or ultrasonographic guidance 35. However, there are some significant drawbacks to this approach. First, the injection of the non-specific fluorophore into the tumor is a critical step in the procedure and has to be very accurate to minimize false-negative and false-positive results. Additionally, spillage/leakage of fluorophore within the mammary gland during the procedure will decrease accuracy of both localization of the tumor and macroscopic margin assessment. Therefore, we believe NIRF-guided surgery should ideally be combined with tumor-targeted fluorophores, which provide molecularly-specific detection of cancer cells. In these agents, the NIR fluorophore has been conjugated to a specific targeting ligand or monoclonal antibody. This allows for tumor-specific binding of the fluorophore, increasing signal-to-noise (SN) ratios and minimizing spillage of the fluorophore during the surgical procedure 3,10. Several studies have shown the feasibility of using tumor-targeted fluorophores in vivo to image tumors intraoperatively, including the use of tumor-targeted ICG-conjugated agents 5,9,22,25,26,36,37. However, there are some significant drawbacks, including the heterogeneity of (breast) tumors which should be solved before applying tumor-targeted NIRF imaging in the clinic. In BCS, the preoperative biopsy taken prior to surgery could provide important information on molecular targets for NIRF imaging. As this biopsy is considered standard practice, it will not require an additional invasive procedure, while offering the possibility to determine the expression of different kinds of molecular targets present on the breast cancer cells by immunohistochemical analysis. The surgeon could then look for NIRF agents suited for each individual tumor, offering a more patient-tailored approach. Conclusion We have preclinically assessed the applicability of NIRF imaging applications in BCS by exploiting tissue-simulating breast phantoms. NIRF-guided intraoperative tumor localization and detection of remnant disease showed feasible. Clinical studies are needed to further validate these results for use in BCS. Conflict of interest statement There are no potential or actual, personal, political, or financial interests by any of the authors in the material, information, or techniques described in the paper. All authors have seen and approved the manuscript and are fully conversant with its contents.. 35. 2.

(36) Acknowledgements This work was supported by a grant from the Jan Kornelis de Cock foundation. The authors wish to thank Mr. T. Buddingh, MD for assistance in operating on the phantoms.. 36.

(37) References 1.. Parkin DM, Bray F, Ferlay J, Pisani P. Global cancer statistics, 2002. CA: a cancer journal for clinicians. 2005;55(2):74-108.. 2. Schwartz GF, Veronesi U, Clough KB, et al. Proceedings of the consensus conference on breast conservation, April 28 to May 1, 2005, Milan, Italy. Cancer. 2006;107(2):242-250. 3.. Pleijhuis RG, Graafland M, de Vries J, Bart J, de Jong JS, van Dam GM. Obtaining adequate surgical margins in breast-conserving therapy for patients with early-stage breast cancer: current modalities and future directions. Annals of surgical oncology. 2009;16(10):2717-2730.. 4.. Singletary S. Breast cancer surgery for the 21st century: the continuing evolution of minimally invasive treatments. Minerva chirurgica. 2006;61(4):333-352.. 5.. Kirsch DG, Dinulescu DM, Miller JB, et al. A spatially and temporally restricted mouse model of soft tissue sarcoma. Nature medicine. 2007;13(8):992-997.. 6.. Luker GD, Luker KE. Optical imaging: current applications and future directions. Journal of Nuclear Medicine. 2008;49(1):1-4.. 7.. Tromberg BJ, Pogue BW, Paulsen KD, Yodh AG, Boas DA, Cerussi AE. Assessing the future of diffuse optical imaging technologies for breast cancer management. Medical physics. 2008;35(6):2443-2451.. 8.. von Burstin J, Eser S, Seidler B, et al. Highly sensitive detection of early-stage pancreatic cancer by multimodal near-infrared molecular imaging in living mice. International journal of cancer. 2008;123(9):2138-2147.. 9.. Trastuzumab complexed to near-infrared fluorophore indocyanine green. Molecular Imaging and Contrast Agent Database (MICAD); 2009. https://www.ncbi.nlm.nih. gov/books/NBK23414/.. 10. Keereweer S, Kerrebijn JD, Van Driel PB, et al. Optical image-guided surgery— where do we stand? Molecular Imaging and Biology. 2011;13(2):199-207. 11. Tagaya N, Yamazaki R, Nakagawa A, et al. Intraoperative identification of sentinel lymph nodes by near-infrared fluorescence imaging in patients with breast cancer. The American Journal of Surgery. 2008;195(6):850-853. 12. Sevick-Muraca EM, Sharma R, Rasmussen JC, et al. Imaging of lymph flow in breast cancer patients after microdose administration of a near-infrared fluorophore: feasibility study 1. Radiology. 2008;246(3):734-741. 13. Brandt MG, Moore CC, Jordan K. Randomized control trial of fluorescence-guided surgical excision of nonmelanotic cutaneous malignancies. Journal of Otolaryngology. 2007;36(3). 14. Ogasawara Y, Ikeda H, Takahashi M, Kawasaki K, Doihara H. Evaluation of breast lymphatic pathways with indocyanine green fluorescence imaging in patients with breast cancer. World journal of surgery. 2008;32(9):1924-1929. 15. Stummer W, Pichlmeier U, Meinel T, et al. Fluorescence-guided surgery with 5-aminolevulinic acid for resection of malignant glioma: a randomised controlled multicentre phase III trial. The lancet oncology. 2006;7(5):392-401. 16. Bremer C, Ntziachristos V, Weissleder R. Optical-based molecular imaging: contrast agents and potential medical applications. European radiology. 2003;13(2):231-243. 17. Ntziachristos V. Fluorescence molecular imaging. Annu. Rev. Biomed. Eng. 2006;8:1-33. 18. Troyan SL, Kianzad V, Gibbs-Strauss SL, et al. The FLARE™ intraoperative near-infrared fluorescence imaging system: a first-in-human clinical trial in breast cancer sentinel lymph node mapping. Annals of surgical oncology. 2009;16(10):2943-2952. 19. Kitai T, Inomoto T, Miwa M, Shikayama T. Fluorescence navigation with indocyanine green for detecting sentinel lymph nodes in breast cancer. Breast cancer. 2005;12(3):211-215. 20. Murawa D, Hirche C, Dresel S, Hünerbein M. Sentinel lymph node biopsy in breast cancer guided by indocyanine green fluorescence. British Journal of Surgery. 2009;96(11):1289-1294. 37. 2.

(38) 21. Gee MS, Upadhyay R, Bergquist H, et al. Human Breast Cancer Tumor Models: Molecular Imaging of Drug Susceptibility and Dosing during HER2/neu-targeted Therapy 1. Radiology. 2008;248(3):925-935. 22. Lee SB, Hassan M, Fisher R, et al. Affibody molecules for in vivo characterization of HER2-positive tumors by near-infrared imaging. Clinical Cancer Research. 2008;14(12):3840-3849. 23. Backer MV, Levashova Z, Patel V, et al. Molecular imaging of VEGF receptors in angiogenic vasculature with single-chain VEGF-based probes. Nature medicine. 2007;13(4):504-509. 24. Chen K, Li Z-B, Wang H, Cai W, Chen X. Dual-modality optical and positron emission tomography imaging of vascular endothelial growth factor receptor on tumor vasculature using quantum dots. European journal of nuclear medicine and molecular imaging. 2008;35(12):2235-2244. 25. Sampath L, Kwon S, Ke S, et al. Dual-labeled trastuzumab-based imaging agent for the detection of human epidermal growth factor receptor 2 overexpression in breast cancer. Journal of Nuclear Medicine. 2007;48(9):1501-1510. 26. Sega EI, Low PS. Tumor detection using folate receptor-targeted imaging agents. Cancer and Metastasis Reviews. 2008;27(4):655-664. 27. Themelis G, Yoo JS, Soh K-S, Schulz R, Ntziachristos V. Real-time intraoperative fluorescence imaging system using light-absorption correction. Journal of biomedical optics. 2009;14(6):064012-064012-064019. 28. Pogue BW, Patterson MS. Review of tissue simulating phantoms for optical spectroscopy, imaging and dosimetry. Journal of biomedical optics. 2006;11(4):041102041102-041116. 29. Alec M, Lomnes SJ, Lee DS, et al. Tissue-like phantoms for near-infrared fluorescence imaging system assessment and the training of surgeons. Journal of biomedical optics. 2006;11(1):014007-014007-014010. 30. Yuan B, Chen N, Zhu Q. Emission and absorption properties of indocyanine green in Intralipid solution. Journal of biomedical optics. 2004;9(3):497-503. 31. Iizuka MN, Sherar MD, Vitkin IA. Optical phantom materials for near infrared laser photocoagulation studies. Lasers in surgery and medicine. 1999;25(2):159-169. 32. Durkin A, Jaikumar S, Richards-Kortum R. Optically dilute, absorbing, and turbid phantoms for fluorescence spectroscopy of homogeneous and inhomogeneous samples. Applied Spectroscopy. 1993;47(12):2114-2121. 33. Wagnières G, Cheng S, Zellweger M, et al. An optical phantom with tissue-like properties in the visible for use in PDT and fluorescence spectroscopy. Physics in medicine and biology. 1997;42(7):1415. 34. Intes X. Time-domain optical mammography SoftScan: Initial Results1. Academic radiology. 2005;12(8):934-947. 35. Berridge D, Mastey L, Eckstrom P, Czarnecki D. Indocyanine green dye as a tissue marker for localization of nonpalpable breast lesions. AJR. American journal of roentgenology. 1995;164(5):1299. 36. Ke S, Wen X, Gurfinkel M, et al. Near-infrared optical imaging of epidermal growth factor receptor in breast cancer xenografts. Cancer research. 2003;63(22):7870-7875. 37. Mieog JSD, Hutteman M, van der Vorst JR, et al. Image-guided tumor resection using real-time near-infrared fluorescence in a syngeneic rat model of primary breast cancer. Breast cancer research and treatment. 2011;128(3):679-689.. 38.

(39) 2. 39.

(40)

(41) CHAPTER 3. Detection of melanoma metastases in resected human lymph nodes by noninvasive multispectral photoacoustic imaging. Sentinel node biopsy in patients with cutaneous melanoma improves staging, provides prognostic information, and leads to an increased survival in node-positive patients. However, frozen section analysis of the sentinel node is not reliable and definitive histopathology evaluation requires days, preventing intraoperative decision-making and immediate therapy. Photoacoustic imaging can evaluate intact lymph nodes, but specificity can be hampered by other absorbers such as hemoglobin. Near infrared multispectral photoacoustic imaging is a new approach that has the potential to selectively detect melanin. The purpose of the present study is to examine the potential of multispectral photoacoustic imaging to identify melanoma metastasis in human lymph nodes. Three metastatic and nine benign lymph nodes from eight melanoma patients were scanned ex vivo using a Vevo LAZR © multispectral photoacoustic imager and were spectrally analyzed per pixel. The results were compared to histopathology as gold standard. The nodal volume could be scanned within 20 minutes. An unmixing procedure was proposed to identify melanoma metastases with multispectral photoacoustic imaging. Ultrasound overlay enabled anatomical correlation. The penetration depth of the photoacoustic signal was up to 2 cm. Multispectral three-dimensional photoacoustic imaging allowed for selective identification of melanoma metastases in human lymph nodes.. International Journal of Biomedical Imaging G.C. Langhout, D.J. Grootendorst, O.E. Nieweg, M.W.M. Wouters, J.A. van der Hage, J. Jose, H. van Boven, W. Steenbergen, S. Manohar, T.J.M. Ruers.

(42) Introduction Sentinel node biopsy in patients with melanoma improves staging and guides the subsequent management of the nodal basin. The tumor status of the sentinel node is the most important prognostic factor. Sentinel node biopsy plus complete node dissection improves ten-year survival in node-positive patients1. Routine histopathology evaluation requires several days, preventing intraoperative decision making and immediate lymph node dissection. Less time consuming procedures like imprint cytology and frozen section analysis are characterized by a low sensitivity. Intraoperative imprint cytology and frozen section show a sensitivity of approximately 30 % and 50 %, respectively 2,3. Also noninvasive imaging techniques like ultrasound, computed tomography (CT), magnetic resonance imaging (MRI), and positron emission tomography (PET) lack the sensitivity to reliably detect micrometastases 4. Photoacoustic imaging is a hybrid imaging technique, combining high-resolution ultrasound with molecular specific optical excitation. The method is based on optical absorption of pulsed light by specific tissue. The absorbed optical energy is converted to heat, which generates ultrasound waves by thermal expansion. Detection of these light induced ultrasound waves allows recovering of the location of the ultrasound sources, which are the tissue structures that absorb the light. In general, structures that absorb more light will cause a stronger ultrasound signal. The absorption of light will differ at various wavelengths, as a molecular property. A multispectral approach makes use of this property to selectively visualize metastases. Melanin is a light absorbing substance and photoacoustic imaging has been shown to detect melanin in lymph node metastases from melanoma using a single wavelength 5,6. Differentiation between blood and melanoma proved to be difficult because both are strong optical absorbers and therefore create a clear photoacoustic signal. Wang used two-wavelength photoacoustic imaging to visualize thin superficial melanomas in mice using both red and infrared light 7. Although the differentiation between melanin and blood was possible, the use of red light limits the penetration depth of this approach. For example, in this study the visualized melanoma was 0.3 mm thick and located at 0.32 mm under the skin of a nude mouse. Human lymph nodes are located at a depth of several centimeters in the human body. Near infrared light penetrates deeper into the tissue. Near infrared multispectral photoacoustic imaging may be able to selectively visualize melanin and blood at centimeters depth in intact human lymph nodes. This study explores the potential of this new approach to identify melanoma metastasis ex vivo in human lymph nodes. A specific purpose was to correlate spectral images with known reference spectra to differentiate melanoma metastases from blood related artifacts in intact human lymph nodes. The other purpose was to investigate the value of a hand-held photoacoustic-ultrasound system capable of 3D imaging.. 42.

(43) Materials and methods Photoacoustic setup The data was acquired using the Vevo LAZR © Photoacoustic Imaging System (FUJIFILM VisualSonics Inc., Toronto), a hybrid photoacousticultrasound reflection mode imager with a 21 MHz (13–24 MHz) linear array transducer. Fiber optic bundles near the surface of the transducer coupled to a 20 Hz tunable laser are able to deliver 20 mJ/cm 2 in the wavelength range of 680–970 nm. For three-dimensional image acquisition, the transducer was automatically moved over the lymph node by a stepping motor with a step size of 0.2 mm. Both ultrasound and photoacoustic images were exported as 16 bit greyscale images to MATLAB (version 7.14; Mathworks, Natick, MA,USA) for image processing. Phantom In order to verify whether multispectral photoacoustic imaging is able to differentiate the chromophores blood from melanin, a phantom was developed. This was made of absorbing and scattering agar gel (2 % in water) mimicking the optical properties of soft tissue. Embedded inside the phantom at 4±1 mm depth were two 2 % agar cylinders (diameter 2 mm, height 6 mm). One cylinder contained bovine hemoglobin (Hb) (SigmaAldrich, Zwijndrecht, the Netherlands) in a concentration of 15 g/L.The second cylinder contained B-16 melanin producing melanoma cells (2 × 10 6 cells/mm 3). The background consisted of 2 % agar completely covering the 6 mm high cylinders with a cover on the top of 4±1 mm. Multispectral volume scanning was performed using five distinct illumination wavelengths between 680 and 840 nm with 40 nm intervals. After image acquisition, the photoacoustic spectrum was obtained by selecting a 3D region of interest in each inclusion and calculating the mean values with standard deviations. Human lymph nodes Twelve human lymph nodes from eight melanoma patients undergoing lymphadenectomy were obtained from the surgical specimen. The experimental protocol was performed according to the Dutch guidelines for clinical research and patient’s informed consent was acquired prior to surgery. The lymph nodes were stored in phosphate buffered saline before and during imaging. Subsequently, histopathology examination was performed using two or more slides stained with hematoxylin and eosin and biomarkers (S100, HMB-45, Melan-A) when needed. In order to obtain accurate reference spectra for both blood and melanin, multispectral images were acquired for each node. Five wavelengths were selected: 3 wavelengths (700 nm, 800 nm, and 820 nm) covered the near infrared range. Two additional near infrared wavelengths (732 nm and 757 nm) were selected based on the absorption spectrum of oxy- and deoxyhemoglobin; both show an increase between 732 nm and 757 nm, whereas the absorption spectrum of melanin shows a slight 43. 3.

No results found